Determining and vigilance of the Road Accidents Hotspots using Machine Learning Algorithms

Abstract

Worldwide, traffic accidents result in fatalities, injuries, and financial losses.Accurate models for predicting accident severity are essential for transportation systems.This study focuses on constructing injury severity classification models using key variables and various machine learning techniques.Supervised algorithms (Random Forests, Decision Trees, Logistic Regression, and K-Nearest Neighbors) are employed, with the SMOTE algorithm addressing data imbalance.Findings indicate that Logistic Regression and SVM models effectively determine injury severity.Additionally, leveraging user GPS data, the system proactively alerts users before reaching accident-prone areas, visually mapping these locations.